Temporary Digital Exhibit by the National Museum of the Great Lakes
Created by Peter Limbert in fulfillment of the BGSU Public History capstone requirement, Spring 2025
Review Committee: Dr. Cheryl Dong, Dr. Savitri Kunze, & Dr. Shirley Green
Unfortunately, as of June 2025, the StoryMaps service has been absorbed into the ArcGIS platform. Therefore, the interactive content element of the Dark Waters project is no longer available.
Contents
Acknowledgements
This project would not have been possible without the tireless support of Carrie Sowden, who had incredible patience and support throughout this project. I would also like to thank Dr. Savitri Kunze, Dr. Shirley Green, Dr. Cheryl Dong, and Dr. Rebecca Mancuso for their contributions to my academic career.
Most importantly, I also thank my parents, Kari and Timothy Limbert. They taught me to value history, the humanities, and the arts. They've supported me in all my endeavors and have made all my accomplishments possible. This project is dedicated to them and to the members of our family who are no longer with us.
Dark Waters: True-Crime, Murder, and Mystery on the Great Lakes is an upcoming temporary exhibition at the National Museum of the Great Lakes. Its mission is to showcase the ways that the Great Lakes have been vehicles for dramatic narratives throughout time, highlighting their role in popular myth and culture. I began working on the project in December of 2023 as a digital exhibit designer & curator. The work that I contributed was used as my capstone project for my MA in History from BGSU. In fulfillment of my Public History certificate program, this page is a summary of the ideas, methods, and processes that went into designing my portions of Dark Waters.
Create an online learning space for audiences unable to visit the museum.
Portray true-crime Great Lakes history as both enjoyable and informative.
Demonstrate the validity of using academic research disciplines in creating popular true-crime content for public consumption.
Raise awareness of the Second Wave expansion.
My contributions to the physical exhibit will feature three interpretive panels I drafted. They cover three distinct topics:
The remembrance of maritime disasters through music
The importance of first responders in the S.S. Eastland catastrophe
The development of forensic analysis in the S.S. Noronic disaster
These brief texts are not meant to give comprehensive knowledge of the subjects they cover, but to draw attention to broader themes in maritime history that are often overlooked. The language is meant to be broad in scope and appeal to emotion in order to encourage audiences to engage in critical thinking rather than just be fed information. The texts are also the limit of my involvement in the development of Dark Waters' physical exhibit.
These panels will be installed along with the launch of the physical exhibit on June 28, 2025.
The digital portion of Dark Waters is the centerpiece to my contributions and my Public History capstone. In January of 2023, I proposed the construction of an entirely new online exhibit space that would harness the latest digital mapping technology in order to tell Great Lakes history in a new and engaging way. I was given approval to work completely independently to build the online exhibit as I saw fit. I managed every stage of research, design, and web construction. Built using the online StoryMaps service, it was a project that changed and evolved over more than a year of development.
The digital portion focused on providing the following core features:
20 distinct historical events between c.1670-1980 placed on an interactive map with relevant data points and content information.
A "zone map" of each Great Lake illustrating what makes them unique for wrecks, weather, trade, and other historical factors.
Visual statistics for the number of estimated shipwrecks in each Lake.
A "Resources" section which links out to a separate webpage that provides information for sources, references, and suggestions for further research.
Spaces for community members to submit their own experiences with Great Lakes mysteries and to leave feedback on the quality of the online exhibit.
The title page of the Dark Waters digital history component, hosted by StoryMaps.
I approached this project knowing I would work on a largely-independent basis due to BGSU's proximity to the research materials offered by the Historical Collection of the Great Lakes at the BGSU Center for Archival Collections, and due to the digital nature of my work not requiring frequent in-person consultation with my supervisor. As a team, we agreed that I would do as much research as possible on a remote basis, using shared folders and portfolios as a way to cache sources, information, content drafts, and progress updates.
The process started with finding a software capable of creating the type of user interface we were looking for. After experimenting with several open-source mapping services, we agreed as a team that the StoryMaps site held the most potential for the tools it offered, its accessibility, and its low threshold for designer training. The main appeal of this service was that it had in-house tools built so that designers wouldn't have to have the technical skills or training to create their own maps using data entry or other complicated processes used by resources like ArcGIS. It also served our needs because we knew that the online exhibit would be temporary by design and not require an enormous amount of data or content integration.
Research began with two driving questions:
What events in Great Lakes history have gone overlooked by the popular media?
How can we tell the stories of these events in a way that appeals to true-crime aesthetics while also using academic disciplines?
The first question would not be answered until several months before the project was finished, as it made up the largest amount of time investment. The second question was answered through our research methodology.
I started by setting parameters. I would find 14 lesser-known, more localized, events that are not commonly discussed in Great Lakes historical discourse. To anchor interest in these stories, I would also find 6 famous events and tell their narratives in unique ways. This amounted to a total of 20 stories. We consulted as a team how many stories we should feature to make the project worth the time, but also not lose audience attention. We agreed that on a minimum-maximum range of 15-25. The end total was the originally-planned 20.
The challenge would be finding the local stories we wanted to emphasize. The natural first place to start was at the Center for Archival Collections on BGSU's campus. After pouring through their collections on Great Lakes history, there was verry little that could be used for the digital history component. There were, however, valuable materials that were used in the construction of the physical exhibit component.
Once the archival route had been exhausted, I turned to online newspaper repositories.
ProQuest offered the highest amount of material that could be publicly accessed, but the image quality of its scans was often low, requiring technical cleanup on the user-end.
Chronicling America offered greater coverage of more widely-reported events, such as disappearances or accidents which received more national news coverage. It also provided higher quality scans of content, requiring very little user-end editing.
Newspaper Archive offered very little that was useful in the research process. Its collections for national publications was high quality, but it often failed to pull sources from local or regional publications, giving us only glimpses at events we wanted to know more about.
By the end of May 2024, I'd compiled a rough list of 15 stories that had enough resources to them to merit their coverage. However, over the next several months the research process grew complicated as I refined my methodology and process.
Searching through these newspaper repositories was often made more challenging depending on the criteria used. We wanted to have at least one event in or near each Great Lake, but this was often difficult to do without literally typing "Lake Michigan" or "Lake Huron" into the search function with another word such as "Disappearance" or "Murder." The required use of specific search queries limited the results being produced, resulting in some lakes having an abundance of events and others very few.
In cases where a particular lake had very few localized news events, I determined it would be best to find a famous event or known shipwreck to compensate. In Lake Ontario, for example, we used the fire of the SS Noronic. In Lake Superior, we harkened to the immortalized Edmund Fitzgerald. This helped us get our desired events-per-lake quota while still including stories that audiences would have at least a passing knowledge of due to their popular culture footprint.
For many stories, it was also challenging to produce scholarly secondary sources. Many of the localized events had no scholarly coverage, or were given minimal reference at best. This resulted in a coverage imbalance where our 6 famous events had an overwhelming amount of source material, while our localized events had comparatively little. This posed issues for our mission of bringing academic research principles into our process. In the end, we determined that the nature of true-crime often hinges upon coverage imbalance. We decided that what made the lesser-covered stories so intriguing was the lack of definitive materials published on them. We decided to downplay the 6 famous events as much as possible and to raise up the localized events in order to remedy this imbalance.
By the end of August, 2024, I'd completed the list of all stories we wanted to include:
Lake Huron’s Ancient Hunting Ground
Lake Michigan’s “Stonehenge”
The Ghost Ship Griffon
The Battle of Lake Erie
Regicide on Beaver Island
An “Unprovoked and Dastardly Crime”
The Frozen Sailor of Malden
“A Romantic Death”
Capsizing of the SS Eastland
Duluth’s Disgrace
Death of the Piattris
The Ghost Ship Kamloops
A Mother-Daughter Pact
The Butcher of Kingsbury Run
Fire of the Noronic
Flight into Thin Air
The Survival of Dennis Hale
The Good Hart Murders
Sinking of the Edmund Fitzgerald
The Slaying of Duane Dilliott
The Frozen Sailor of Malden was the first local story I discovered through my research. It was the exciting find that confirmed the value of the project.
The Chronicling America repository's results for the murder of George Northup, one of our featured local stories.
Partial list of stories as they appear in the online exhibit.
I started designing the initial StoryMap very early on in the development stage. One of the first things I had to decide was which map function was best suited for our needs. The site offered several types of maps, including data-driven maps, statistical maps, and tour-style maps, of which I chose the last. Tour-style worked in our favor because many of the local stories we wanted to tell didn't have much visual media to display. In most instances, we only had newspaper clippings, which would not have looked very impressive. I knew from discussions with our team that we wanted a map that showed users the entire Lakes region, so the Explorer List design was eventually chosen (see image carousel).
Using this template, I arranged each event chronologically and gave them a pin on the regional map. For most of our featured stories, this wasn't difficult. Press were almost always very descriptive of locations, sometimes even down to the street level. However, in some instances it became clear that using a single point would be problematic. In the disappearance of Le Griffon, for example, it was difficult to choose a specific single point because the ship has never been found. In the end, me decided as a team to place the marker at the entrance to Green Bay to mark the general vicinity experts believe it may rest. This was a compromise that suited us so long as we made this distinction clear in the text.
I knew from the very beginning that it was essential to get the look nailed down quickly. After all, a good look would anchor the whole project. Initially, the design I chose used very dark colors (black background with neon blue, green, and white text) to give it a true-crime aesthetic. As time went on, however, it became clear that this didn't fit the tone we wanted. In February of 2025, I was given access to the NMGL brand colors and fonts, which were implemented over a weeklong process of experimenting with StoryMaps' template design feature.
I ran into numerous frustrations with the template designer. StoryMaps ran using the Google Fonts tool, which by default only allows access to certain fonts that are either owned by Google or are in the public domain. This meant that the specific fonts used by the NMGL weren't available in StoryMaps, and there was no way to add them manually. The way around this obstacle was to find fonts that were close enough in appearance to mimic the actual fonts. Instead of Palatino Linotype, I used PT Serif Pro. Fortunately, Open Sans was available in StoryMaps. This was fortunate since it was going to be the main type font used for the map content.
StoryMaps was also problematic because of the limited ability to incorporate multiple theme colors. It allowed for up to three colors, but they were locked into specific functions. A header, for example, could only be one choice of color and the background had to be another. If the designer registered two colors as being too similar, it would often refuse to allow them to be added in the same theme package. To work around this, I experimented back and forth in the designer for three days to get the colors just right enough for them to be allowed. I chose the NMGL Logo Yellow for titles, Light Blue for accents and functions, and white for body text. The background was given Dark Blue to contrast with the others. This allowed the main text to be easily legible while giving the page much-needed variety for the eye.
Part of the designers' options for adjusting colors and typography.
Figuring out how I wanted to have the entire page designed was more difficult, and was always the point where I asked myself, "Is this right? Do I like this?". What made it challenging was that StoryMaps would insert a map or text segments in a single column; it did not allow for elements to be placed side-by-side. I struggled with figuring out how I wanted the page to be laid out, inspiring me to look to other projects for reference.
The best model I found was published by the Michigan History Center in 2024. It differed from Dark Waters because it had maps built using ArcGIS StoryMaps rather than just StoryMaps, but the design philosophies were the same. I took detailed notes on how they organized their page, divided space, and made room for each content segment to stand on its own while also being balanced by map segments. Throughout development of Dark Waters, I would find myself frequently going back and asking myself new questions such as "How did the MHC arrange their content?" or "What can I do to improve on the design they used?" That last question became more prominent as Dark Waters gained steam and grew in size.
I wanted there to be as little interpretive text outside of maps as possible. The point was to get audiences engaged with the mapping tools, not lecture them. I wanted especially to avoid the classic museum pitfall of putting too much text and too little interactive substance. The image to the right is all the introductory text I drafted to outline the goal, theme, and instructions. With minimal text, audiences are encouraged to dive in (pun intended) to the content as quickly as possible.
To break up large media segments, I used brief amounts of text. This meant audiences weren't constantly barraged by things to do themselves; it provided breathing space. The text also helped to explain some of the media, giving vital context about why certain information was presented the way it was.
Because much of maritime history revolves around the use of statistics (# of shipwrecks, storms, lake dimensions, etc.) to emphasize the importance of qualitative information, I wanted to find a way that could synthetize large amounts of data into a visually appealing format while still being easy to understand. When it came to choosing which set of data to display, our team and I agreed that for most audiences, what would be most in-theme for Dark Waters would be displays of how many ships have been lost and how many have been found in each lake.
Before we could visualize this information, however, we had to conduct more research to determine ballpark estimates of known shipwrecks. There is no single source that compiles a complete accounting of all known wrecks in the Lakes, so information was drawn from sources such as the Michigan Historical Center, the Minnesota Historical Society, the Ohio History Connection, Nature Canada, NASA, and NOAA. No one knows exactly how many ships have been lost, with estimates in each lake ranging from a few hundred to several thousand. We took the numbers which appeared most frequently and which were most realistic to create a bar graph depicting the information in 2 formats:
Shipwrecks that have been FOUND in each Lake.
Shipwrecks that have been ESTIMATED to have sunk in each Lake.
All numbers involved are estimates, which we made sure to emphasize both in the graph labels and interpretive text. We also used the opportunity to emphasize how new discoveries are being made as time goes on, which will hopefully inspire further audience engagement with various maritime history resources.
"Wrecks in the Great Lakes" graph, built using StoryMaps' graphing features.
To showcase the work that went into building the digital Dark Waters, I created a second web space where audiences could find all the resources used in our research. This Resources page was hosted on a Google Sites page that would be open to the public, featuring an internal page for each of the 20 events covered in the StoryMap.
This space was designed to be much less visually and technically complex than the StoryMap, offering immediate access to all the raw information and sources for the user to engage with at their leisure. Each event covered in the StoryMap had a page where we included links to news articles, scholarly literature, historical society websites, archive repositories, and the images used in the exhibit. This was not only a way to show our work, but also to allow audiences a way to do their own preliminary research based off the information we provided. It gave them the keys to go down the rabbit hole themselves, as it were.
Included in the Resources page were also links to relevant articles on subjects found in the NMGL's Inland Seas journal, the longest-running scholarly publication focused on Great Lakes history. Published in 1944, its articles cover topics such as marine archaeology, cultural history, industrial history, and others. In a revolutionary expansion of access, over 300 Inland Seas articles published between 1944-2020 have been digitized and added to an online repository where they can be searched using textual parameters and viewed in their entirety.
Where relevant, events on the Resources page have links directing audiences to Inland Seas issues, giving them more inspiration to interact with NMGL materials. This access also encourages monetary or volunteer contributions to the museum, reflecting the need for greater community engagement to support museum efforts.
Since Dark Waters is designed to appeal to the popular imagination of the general public, successful audience engagement is crucial. From the start of the design process, our team wanted to find a way to bring out the ways in which every person is connected to a sense of wonder and mystery. The NMGL highlights how the Great Lakes mean so many different things to different people, so we wanted a way for people to share their own unique mystery stories with us.
The most practical solution we came up with was to embed a public response questionnaire using the Google Forms service. We would insert the form into the StoryMaps page with 2 spaces for audience members to type responses.
Space 1 was where they could submit their own experiences with Great Lakes mystery.
Space 2 was an optional space where they could type their contact information in case they wanted us to contact them about their story.
Even when we first started discussing this option, we knew that moderation would need to be key. We linked the Google form to a secondary email address owned by the museum so any potential submissions wouldn't obstruct any active workflows. Using Google Forms also allowed us to avoid AI bot submissions.
We also wanted our audience to be able to leave us feedback on what parts of our exhibit worked and what didn't. To this end, we included a second Google Form for this purpose.
In late March, StoryMaps announced that it was discontinuing its services, effective May 31, 2025. This announcement meant that the end-of-life date for Dark Water's online component would be moved up considerably.
As a team, we had several long discussions about what we would like to do with the exhibit before it goes offline. One solution was to register it for migration to ArcGIS StoryMaps for an institutional fee that was well above our original funding predictions. Another was to get a temporary extension to last until July 26, giving us at least several months of use while the physical temporary exhibit was open.
We have yet to settle on a solution to this challenge, as our discussions for implementing Dark Waters are ongoing.
In our original plan, we wanted to include a second StoryMap featuring timelines and stories related to the Licavoli crime family in Toledo during the 1930s. The goal of this second map was to provide Dark Waters with a local history connection and highlight the organized criminal element of Great Lakes history.
Development on this second map took place alongside Dark Waters. It included a scan of a 1936 Toledo street map with pins for notable events, character sheets for the 3 heads of the family, an interactive map on Great Lakes smuggling, and a linear timeline of gang activities with images.
In March of 2025, I made the executive decision to halt progress on the Licavoli map. Because of time constraints, source availability, and service migration issues, it was deemed too large a project to finish in a way that I felt would be satisfactory. It was a difficult decision, but it allowed more time and resources to be dedicated to making Dark Waters a fully fleshed-out and polished product. Because of StoryMaps' service retirement, it is unlikely this project will ever be made available to the public in its unfinished state.
